| title |
Quickstart |
| description |
Start tracing your LLM calls in under 5 minutes |
## Installation
```bash
pip install fallom
# With auto-instrumentation for your LLM provider:
pip install fallom opentelemetry-instrumentation-openai
pip install fallom opentelemetry-instrumentation-anthropic
```
## Quick Start
<Warning>
**Import order matters!** You must import and initialize Fallom **before** importing OpenAI or other LLM libraries.
</Warning>
```python
import fallom
fallom.init(api_key="your-api-key")
# NOW import OpenAI (after instrumentation is set up)
from openai import OpenAI
client = OpenAI()
# Set default session context for tracing
fallom.trace.set_session("my-agent", session_id)
# All LLM calls are now automatically traced!
response = client.chat.completions.create(
model="gpt-4o",
messages=[{"role": "user", "content": "Hello!"}]
)
```
```bash
npm install @fallom/trace
```
## Quick Start
```typescript
import fallom from "@fallom/trace";
import OpenAI from "openai";
// Initialize Fallom
await fallom.init({ apiKey: "your-api-key" });
// Wrap your LLM client for automatic tracing
const openai = fallom.trace.wrapOpenAI(new OpenAI());
// Set session context
fallom.trace.setSession("my-agent", sessionId);
// All LLM calls are now automatically traced!
const response = await openai.chat.completions.create({
model: "gpt-4o",
messages: [{ role: "user", content: "Hello!" }],
});
```
Every LLM call is automatically captured with:
| Field |
Description |
| Model |
The LLM model used (e.g., gpt-4o, claude-3-opus) |
| Tokens |
Input and output token counts |
| Latency |
Request duration in milliseconds |
| Prompts |
Full prompt messages sent to the model |
| Completions |
Model responses |
| Session |
Your config key and session ID for grouping |
Create a free account at [bench.fallom.com](https://bench.fallom.com)
Set up your first project in the [dashboard](https://bench.fallom.com)
Find your API key in project settings
Get your API key and view your traces
Test different models to optimize cost and performance.
Experiment with prompts to improve outputs.